Neurocomputing is computing using living neurons. This research thrust has the potential to not only shed insights into the underlying cognitive processes in our brain by providing a programmable substrate but to be used to gain insights into better ways of designed spiking neural networks, neuromorphic computing hardware, and AI more generally. Furthermore, the Mind in Vitro project hopes to use bio-computing systems which are extremely energy efficient to replace some power intensive implementations, provide novel computational power, and be used as the 'brain' in bio-hybrid robotics.

Neurocomputers for Neuroscience

The infrastructure developed in the making of a neurocomputer allows for an ideal in vitro neuroscience test-bench in which inputs may be given and outputs may be read under specific parameterizations of the network (e.g. structure, cell concentrations, media, etc) to study properties at the individual neuron and network level.

"What I cannot create, I do not understand"

-Richard Feynman

approaches to record neural activity allow us to visualize the activity of thousands of neurons. However, we do not know the language by which these patterns of neural activity are interpreted, afforded meaning, or translated into appropriate behavioral or cognitive output. Put simply, we do not have a logic for the transformation of neural activity into thought and action. I view discerning a logic as the most important future direction of neuroscience.

-Richard Axel

Following the above quotes by two Nobel Laureate Richards, by creating these brain-like systems, we hope to gain a more detailed understanding of the functioning of our own brains, and by using a programmable substrate, we hope to better understand the transformation of neural activity into thought and action. Check out the work being done below with the two NSF funded projects.

Mind in Vitro -- Computing with Living Neurons • Synthetic Neurocomputers for Cognitive Information Processing